Expected Value (EV) in Sports Betting: The One Number That Matters

Ask a casual bettor whether a wager was good and they will tell you whether it won. Ask a sharp one and they will tell you the price. That gap is the whole game. Winning a single bet tells you almost nothing about whether it was a smart bet, because a coin can land heads five times in a row. Expected value, or EV, is the number that separates a good decision from a good result. It is the one metric that, over enough bets, actually predicts whether you make money. This guide builds it from scratch: how odds hide a probability, how to compare that against your own estimate, how the sportsbook's margin quietly taxes you, and how to run the math on a real line.

What expected value actually means

Expected value is the average amount you would win or lose per bet if you could place that exact bet an unlimited number of times. It is a weighted average: each possible outcome multiplied by its probability, all added together. If the number is positive, the bet pays you more than it should on average and is worth making. If it is negative, you are handing the book an edge. The formula for a single wager is simple:

The hard part is never the arithmetic. It is the probabilities. The book gives you one probability baked into the odds. You have to supply the other one, and you have to be honest about it. Overrate your own read and every bet looks +EV when it is not.

Step 1: turn odds into an implied probability

Every set of odds is a probability in disguise. American odds convert like this. For a favorite priced at negative odds, implied probability = risk / (risk + 100), so -150 becomes 150 / 250 = 60%. For an underdog at positive odds, implied probability = 100 / (odds + 100), so +130 becomes 100 / 230 = 43.5%. Decimal odds are even cleaner: implied probability is just 1 divided by the decimal price, so 2.50 means 1 / 2.50 = 40%. This implied number is the market's stated opinion of how likely the outcome is, plus the book's cut. Getting comfortable with this conversion is the foundation, and a tool like the Statly odds converter does it instantly for any format.

Step 2: the vig, and why implied probabilities lie

Here is the catch that trips up beginners. Add up the implied probabilities on both sides of a market and they come to more than 100%. On a standard -110 / -110 point spread, each side implies 52.4%, and 52.4% + 52.4% = 104.8%. Those extra 4.8 percentage points are the vig, also called the juice or the hold. It is the book's margin, the price of doing business, and it is the single biggest reason most bettors lose slowly over time. The raw implied probability from a line is always inflated by this margin, which means it overstates how likely the book truly thinks the outcome is.

To find the book's real, fair estimate you strip the vig out. The quickest method is to divide each side's implied probability by the total of both sides. In the -110 example, 52.4% / 104.8% gives a no-vig probability of exactly 50%, which is what we would expect from a true pick-em. These vig-free numbers are your honest benchmark. Comparing your estimate against a no-vig line rather than a raw one is one of the most important habits a serious bettor can build. The Statly no-vig calculator handles the de-juicing for two-way and three-way markets automatically.

Step 3: supply your own probability

This is where the skill lives. The book gives you an implied number. Your job is to independently estimate the true probability and see whether it disagrees. There is no single right method, but the disciplined ones share a trait: they are built on data and repeatable process, not gut feel. A few honest approaches:

• Historical hit rate. How often has this exact outcome actually happened in a large, relevant sample? If a player has cleared this rushing line in 7 of his last 10 games against similar defenses, that is a real starting probability, not a hunch.
• Base rates and matchup adjustments. Start from a long-run average, then adjust for the specific matchup, pace, rest, weather for outdoor games, and role. Small, evidence-backed adjustments beat big emotional ones.
• A model. A projection system that runs every game through the same factors removes the recency bias and favorite-team bias that wreck human estimates. This is exactly what the Statly Edge model is built to do, and its Model Builder lets you weight the factors yourself.
• Market consensus as a sanity check. If your estimate is wildly different from the sharp no-vig line at several books, assume you are wrong until you can explain precisely why. The market is not always right, but it is right far more often than a bettor guessing on a Tuesday.

Be ruthlessly honest here. The most common way bettors lose is not bad math, it is inflated probability estimates on the teams and players they already like. If you cannot articulate why your number differs from the market's, you probably do not have an edge.

A full worked example

Let's price a real-shaped bet end to end. Suppose a book lists a quarterback to throw over 1.5 touchdown passes at +115, and the under 1.5 at -145. You want to know whether the over is a +EV bet at a $100 stake.

Notice what did the work. The plus-money odds did not make this a good bet. The disagreement between your probability and the vig-free market probability did. Same price, different estimate, opposite conclusion. You can run this exact calculation on any line with the Statly EV calculator instead of doing it by hand.

Why +EV over volume beats picking winners

Here is the mindset shift that matters most. A +EV bettor fully expects to lose plenty of individual bets. In the example above, the over hits only half the time. You will lose it constantly. That is fine, because on the times it wins you are paid more than the fair price, and across hundreds of similar bets the small edge compounds into a real return. This is the same logic that runs a casino: the house edge on a single hand is tiny and any one player can walk away up, but over millions of hands the math is inescapable. As a bettor, you want to be the house.

The flip side is that variance is genuinely punishing over small samples. A bettor with a legitimate 3% edge can still be down after 200 bets purely from bad luck. This is not a flaw in the approach, it is a feature of any probabilistic edge, and it is exactly why bankroll management and staking discipline matter as much as the picks themselves. Bet a consistent, sensible fraction of your bankroll so that a normal losing streak cannot bust you before your edge has room to play out. Flat staking or a conservative Kelly fraction are both reasonable frameworks here.

Putting it into practice

1. 1Convert the line you are looking at into an implied probability.
2. 2Strip out the vig using both sides of the market to get the fair, no-vig probability.
3. 3Estimate the true probability yourself, from data and process, not from which team you like.
4. 4Run the EV formula. Only bet when your honest estimate beats the no-vig number by a meaningful margin.
5. 5Size the bet as a small, consistent slice of your bankroll and repeat the process across many bets, expecting variance along the way.

Do that consistently and you stop asking whether a bet won and start asking whether it was priced correctly. That single change in the question you ask is what separates gamblers from analysts.